python - sk-learn saved model to disk, but get only array -


when storing fitted_clf sk-learn classifier like:

joblib.dump(fitted_clf, some_path) 

most of time when loading memory like:

joblib.load(some_path) 

only array of array(['col1', 'col2], dtype=object) returned instead of loading fitted pipeline.

however, real pipeline, not understand why not consistent behavior.

edit

i think has different joblib verisions i.e. sklearn.externals import joblib works, when using regular joblib array

confirmed. using sklearn.externals import joblib fixing have consistent behavior.


Comments